Interventional Radiology 937

Chemosaturation with Percutaneous Hepatic Perfusions of for Hepatic Metastases: Experience from Two European Centers Chemosaturation mittels perkutaner hepatischer Perfusion von Melphalan zur Behandlung von Lebermetastasen: Erfahrungen von zwei europäischen Zentren

Authors T. J. Vogl1, S. Zangos1, J. E. Scholtz1, F. Schmitt1, S. Paetzold2, J. Trojan3,F.Orsi4, G. Lotz5, P. Ferrucci6

Affiliations Affiliation addresses are listed at the end of the article.

Key words Zusammenfassung Abstract ●" melphalan ! ! ●" liver Ziel: Chemosaturation mit perkutaner hepatischer Purpose: Chemosaturation with percutaneous ●" metastases Perfusion (PHP; Hepatic CHEMOSAT® Delivery hepatic perfusion (PHP; Hepatic CHEMOSAT® De- ●" hepatic perfusion System; Delcath Systems Inc, USA) ist ein minima- livery System; Delcath Systems Inc, USA) is a ●" hemofiltration ●" safety linvasives, wiederholbares regionales Therapie- minimally invasive, repeatable regional therapy verfahren bei nicht resezierbaren Lebermetas- for unresectable hepatic metastases. It uses a sys- tasen, das mithilfe eines Systems aus Kathetern tem of catheters and filters to isolate hepatic ve- und Filtern das venöse hepatische Blut vom nous blood from the systemic circulation, allow- Systemkreislauf isoliert und eine hochdosierte ing delivery of high-dose to the Chemotherapeutikagabe über die hepatischen Ar- hepatic artery. Effluent hepatic venous blood is terien ermöglicht. Der venöse hepatische Abfluss filtered before being returned to the systemic cir- wird zur Reduktion von Nebenwirkungen des culation, thereby reducing exposure to chemo- Chemotherapeutikums vor dem Rückführen in therapy. We describe our experiences with che- den systemischen Kreislauf gefiltert. Wir berich- mosaturation-PHP at 2 European centers. ten über unsere Erfahrungen mit Chemosatura- Materials and Methods: 14 patients presented un- tion-PHP an zwei europäischen Zentren. resectable hepatic metastases from solid tumors; Material und Methoden: 14 Patienten zeigten nicht 13 received 1 – 3 sessions of chemosaturation- resezierbare hepatische Metastasen solider Tumo- PHP. Melphalan 2.0 (n = 1) or 3.0 (n = 12) mg/kg re; 13 Patienten wurden in 1 – 3 Therapiesitzungen was given as a 30-minute infusion into the hepa- mittels Chemosaturation-PHP behandelt. Melpha- tic artery. 12 patients were evaluable for tumor lan 2,0 (n = 1) und 3,0 (n = 12) mg/kg wurde 30 response. received 20.2.2013 accepted 26.12.2013 Minuten in die hepatischen Arterien infundiert. Results: One complete (cholangiocarcinoma, n = 1) Für das Tumoransprechen waren 12 Patienten aus- and 6 partial responses (ocular, n = 3 or cutaneous Bibliography wertbar. , n = 3) were observed, 5 patients had DOI http://dx.doi.org/ Ergebnisse: Komplette Remission wurde in einem stable disease (ocular melanoma, n = 3; breast can- 10.1055/s-0034-1366081 Patienten beobachtet (Cholangiokarzinom, n = 1), cer, n = 1; gastric , n = 1). Mild to moderate This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited. Published online: 11.4.2014 Fortschr Röntgenstr 2014; 186: partielle Remission in 6 Patienten (Aderhautmela- filter-related toxicity (i. e. thrombocytopenia, ane- 937–944 © Georg Thieme nom, n = 3; malignes Melanom, n = 3), 5 Patienten mia) was observed immediately post-procedure. Verlag KG Stuttgart · New York · zeigten stable disease (Aderhautmelanom, n = 3; Grade 3/4 melphalan-related pancytopenia devel- ISSN 1438-9029 Brustkrebs, n = 1; Magenkarzinom, n = 1). Milde oped after 1 – 2 weeks. All hematological events bis mäßige filterassoziierte Nebenwirkungen (z. B. were managed effectively with transfusions and/ Correspondence Prof. Thomas J. Vogl Thrombozytopenie, Anämie) wurden unmittelbar or other supportive measures. The new high-effi- Institut für Diagnostische und nach der Behandlung beobachtet. Grad 3/4 Mel- ciency filter showed milder toxicity and faster re- Interventionelle Radiologie, phalan-assoziierte Panzytopenien entwickelten covery. In one case, chemosaturation-PHP was J. W. Goethe-Universität sich nach 1 – 2 Wochen. Alle hämatologischen abandoned prematurely due to heparin-induced Frankfurt Ereignisse wurden effektiv mit Transfusionen und/ vaginal bleeding, and one patient died due to ret- Theodor-Stern Kai 7 oder anderen unterstützenden Maßnahmen be- roperitoneal hemorrhage from heparin anti-coag- 60596 Frankfurt handelt. Mit dem neuen, hocheffizienten Filtersys- ulation. Germany Tel.: ++ 49/69/63 01 72 77 tem gab es geringere Nebenwirkungen und eine Conclusion: Chemosaturation-PHP for non-re- Fax: ++ 49/69/63 01 72 58 beschleunigte Erholung. In einem Fall musste das sectable liver metastases is a feasible treatment [email protected] Verfahren aufgrund von einer heparininduzierten option when performed by an experienced mul-

Vogl TJ et al. Chemosaturation with Percutaneous … Fortschr Röntgenstr 2014; 186: 937–944 938 Interventional Radiology

vaginalen Blutung vorzeitig abgebrochen werden und ein Patient ti-disciplinary team. It may be a promising regional therapy for starb aufgrund retroperitonealer Blutung unter Heparinantikoagu- patients with no effective treatment options. lation. Citation Format: Schlussfolgerung: Chemosaturation-PHP nicht resektabler Leber- ▶ Vogl TJ, Zangos S, Scholtz JE et al. Chemosaturation with Per- metastasen ist eine geeignete Behandlungsoption, die von einem cutaneous Hepatic Perfusions of Melphalan for Hepatic Me- erfahrenen, multidisziplinären Team durchgeführt werden kann. tastases: Experience from Two European Centers. Fortschr Es scheint ein aussichtsreiches regionales Verfahren für Patienten Röntgenstr 2014; 186: 937–944 ohne andere effektive Behandlungsmöglichkeiten zu sein.

Introduction phalan is widely available and relatively inexpensive, making it ! an accessible choice for clinics around the world. The liver has a unique dual blood supply which makes regional The purpose of the present article is to describe our experiences treatment possible. Whereas normal hepatocytes receive their with the Hepatic CHEMOSAT® Delivery System in patients with blood primarily from the portal vein, liver tumors are supplied unresectable hepatic metastases treated at two European cen- almost exclusively (up to 95 %) by the hepatic artery [1]. This al- ters. It describes the first patients treated in Europe with this de- lows isolation of the hepatic arterial inflow and venous outflow, livery system. In addition, a second-generation high-efficiency and selective delivery of cytotoxic drugs to unresectable liver me- filter was approved for use in conjunction with the Hepatic CHE- tastases while sparing healthy liver tissue. Regional chemother- MOSAT® Delivery System in Europe in 2012. The phase I, II and III apy procedures include hepatic arterial infusion, percutaneous clinical trials were performed using the first-generation filter, so hepatic perfusion, transarterial (chemo)embolization (TACE) the present report includes the first clinical data with the new and selective internal radiation therapy [2]. filter. Chemosaturation with percutaneous hepatic perfusion (chemo- saturation-PHP) has been developed as a minimally invasive and repeatable regional therapy. It relies on placing a unique double- Patients and Methods balloon catheter percutaneously into the inferior vena cava to ! isolate the hepatic venous blood. High doses of chemotherapy Patients can then be infused directly into the hepatic artery. A fenestrated Between January 2012 and February 2013, 14 consecutive pa- section in the double-balloon catheter allows the isolated hepatic tients with unresectable hepatic metastases from various solid blood to be filtered extra-corporeally before being returned to tumors underwent chemosaturation-PHP with melphalan at the the systemic circulation. The feasibility of chemosaturation-PHP Frankfurt University Hospital (n = 7) and the European Institute has been shown in several studies of patients with unresectable of in Milan (n = 7). Both centers adhered to the Hepatic hepatic metastases or primary hepatic cancer [4 – 9]. CHEMOSAT® Delivery System product instructions with regard to Chemosaturation-PHP has been developed commercially (Hepa- contraindications and precautions. Data evaluation was per- tic CHEMOSAT® Delivery System; Delcath Systems Inc., New York, formed retrospectively. NY) to make the procedure simpler and more widely accessible. A formal clinical trial program for the Hepatic CHEMOSAT® Deliv- Pre-procedural assessments ery System is ongoing; phase I [9], II [10, 11] and III studies [12, Prior to treatment, a physical examination, laboratory tests, and 13] have recently been completed. The phase I study established comprehensive imaging including computed tomography (CT) of that melphalan 3.0 mg/kg was the maximum tolerated dose deli- the thorax and abdomen, upper abdominal magnetic resonance verable by chemosaturation-PHP [9]. An overall response rate imaging (MRI) and, if clinically indicated, a brain MRI, bone scin- (i. e., complete plus partial response) of 50 % was also documen- tigraphy or positron emission tomography (PET) scan were per- ted in patients with metastatic ocular melanoma [9]. These favor- formed. The day before (Frankfurt) or the same day (Milan) of able results prompted a randomized multicenter phase III trial treatment, a complete visceral angiogram was performed to comparing chemosaturation-PHP delivery of melphalan with identify potential variant vascular anatomy. Embolization of se- best alternative care (BAC) in 93 patients with unresectable hepa- lected arterial branches supplying the gastrointestinal tract was

tic metastases from ocular or cutaneous melanoma. When com- performed as needed to avoid inadvertent administration of che- This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited. pared with BAC, chemosaturation-PHP was associated with a sig- motherapy into gastrointestinal or visceral arterial branches. nificant 6.5-month improvement in hepatic progression-free survival, the primary study endpoint (median 8.1 vs. 1.6 months Treatment with BAC; hazard ratio 0.34; p < 0.0001) [13]. Patients received melphalan delivered using the Hepatic CHE- Melphalan was selected as the chemotherapeutic agent for the MOSAT® Delivery System (●" Fig. 1). The procedure was per- formal clinical trial program of the Hepatic CHEMOSAT® Delivery formed under general anesthesia in an interventional radiology System on the basis of several observations. Firstly, it does not suite. Melphalan was given at a dose of 3.0 mg/kg ideal body cause significant liver toxicity even when given at myeloablative weight (maximum 220 mg/treatment) as a 30–minute infusion doses [14, 15]. Secondly, melphalan delivered by operative isolat- into the hepatic artery. Venous effluent blood was filtered via ed hepatic perfusion has previously shown efficacy in patients the extracorporeal hemofiltration circuit during and for 30 min- with hepatic metastases from a variety of , including mel- utes after each infusion. Percutaneous venous access was per- anoma [16 – 19], colorectal cancer [20 – 22], hepatocellular carci- formed under ultrasound guidance in order to reduce the num- noma [23], and neuroendocrine tumors [24]. These data show ber of puncture attempts and any possible cause of bleeding. that the melphalan doses deliverable by hepatic perfusion are Heparin (400 IU/kg body weight) was administered during the adequate for efficacy against a range of solid tumors. Lastly, mel- procedure to ensure free extracorporeal flow and filtration.

Vogl TJ et al. Chemosaturation with Percutaneous … Fortschr Röntgenstr 2014; 186: 937–944 Interventional Radiology 939

Fig. 1 Schematic diagram of the chemosatura- tion-PHP delivery system (Hepatic CHEMOSAT® De- livery System; Delcath Systems Inc., New York, NY) which consists of a closed circuit of catheters and filters designed to deliver chemotherapy to the hepatic artery and then filter effluent hepatic venous blood before it is returned to the systemic circulation.

Abb. 1 Schematische Darstellung des Chemosa- turations-PHP Zuführsystems (Hepatic CHEMOSAT Hepatic CHEMOSAT® Delivery System; Delcath Sys- tems Inc., New York, NY) bestehend aus einem geschlossenen Kreislauf aus Kathetern und Filtern, welches der Infusion der Chemotherapie in die Le- berarterie dient sowie das hepatisch venöse Blut fil- tert, bevor es in den systemischen Kreislauf zurück- geführt wird.

Response and toxicity Table 1 Baseline characteristics. CT, MRI and/or PET scans of the liver were performed at 4- to 8- Tab. 1 Angaben zu Behandlung und Patienten. week intervals. Tumor response of liver lesions was assessed using RECIST criteria [25]. Systemic and local adverse events Frankfurt Milan total were classified by the National Cancer Institute Common Termi- (n = 7) (n = 7) (n = 14) nology Criteria for Adverse Events, version 3.0. Only systemic gender, n events (which did not normalize within 24 hours) and elevations male 4 3 7 of hepatic transaminases (which did not normalize within 7 female 3 4 7 days) were reported. age, years median 53 54 54 range 43 – 64 43 – 74 43 – 74 Results primary tumor, n ! ocular melanoma 5 3 8 cutaneous melanoma 1 2 3 Patients gastric cancer 0 1 1 Patient demographics and tumor characteristics are shown in breast cancer 1 0 1 " ● Table 1. Patients had ocular (n = 8) or cutaneous melanoma cholangiocarcinoma 0 1 1 (n = 3), breast cancer (n = 1), gastric cancer (n = 1) and cholangio- extrahepatic disease, n 1 0 1 carcinoma (n = 1). All patients, except for 1, had metastases con- eastern cooperative oncology group status fined to the liver. All patients had disease progression despite re- 07714 ceiving a range of other treatments for hepatic metastases. karnofsky index 100 % 7 7 14 Chemosaturation-PHP child-pugh score

13 patients received treatment; melphalan was not adminis- a7714This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited. tered in 1 patient because of vaginal bleeding. Patients treated prior therapy for hepatic metastases, n in Frankfurt received a single chemosaturation-PHP treatment, transarterial chemoembolization 4 1 5 systemic chemotherapy 4 6 10 and patients treated in Milan received 1 – 3 treatments. A total hepatic resection 3 1 4 of 18 (Frankfurt, n = 6; Milan n = 12) chemosaturation-PHP pro- microwave ablation 1 0 1 cedures were performed. The interval between repeat treat- selective internal radiotherapy 1 0 1 – ments ranged from 57 177 days. The recommended melpha- radiofrequency ablation 1 1 2 lan dose of 3 mg/kg was given in all patients but one who received a 2 mg/kg dose because of aberrant hepatic vasculari- zation. Procedural care The first-generation filter only was used in 3 patients (Frank- Frankfurt experience furt, n = 2; Milan n = 1), the second-generation filter only was Preprocedural embolization was necessary in 6 patients because used in 7 patients (Frankfurt, n = 4, Milan, n = 3), and 3 patients gastroduodenal artery (GDA) branches were close to the proper (Milan, n = 3) were treated using the first-generation filter for hepatic artery. In one patient with metastases in both liver lobes, their first treatment and the second-generation filter for repeat chemotherapy was injected into the right hepatic artery rather treatment(s). than the proper hepatic artery to avoid unwanted infusion of

Vogl TJ et al. Chemosaturation with Percutaneous … Fortschr Röntgenstr 2014; 186: 937–944 940 Interventional Radiology

the gastrointestinal vessels. One patient developed hepatic arter- Milan experience ial spasm during the angiogram which was treated successfully After treatment, all patients were admitted to the ICU for an aver- with intra-arterial glyceryl trinitrate 0.6 mg. On inflation of the age of 12 hours, except for one patient who died in the ICU 30 occlusion balloons and establishment of the hemofiltration circu- hours post-treatment. There was no difference between the first- lation, transient hypotension was seen in all patients. The median and second-generation filters in terms of hospital stay (average total procedure time was 225 (range, 145 – 270) minutes. The 5.5 days). average venovenous bypass time was 74 (range, 68 – 81) minutes which included positioning and inflation of the occlusion bal- Tumor response loons, the melphalan infusion and washout period. 1 complete response (Milan, n = 1) and 6 partial responses (Frankfurt, n = 2; Milan, n = 4) according to RECIST criteria were Milan experience observed. Stable disease was documented in a further 5 patients Preprocedural embolization was performed in 6 patients. GDA (Frankfurt, n = 4; Milan, n = 1). 2 patients were not evaluable for coiling was necessary in 2 patients because the proper hepatic ar- tumor response (procedure abandoned because of vaginal bleed- tery was very short. Right gastric artery embolization was also ing, Frankfurt, n = 1; patient died shortly after treatment, Milan, necessary in 1 of these patients. 3 patients underwent emboliza- n = 1). tion of the right phrenic arteries. In all patients but one, superse- lective drug injection was achieved by separately cannulating the Frankfurt experience left and right arteries. One of the patients in whom the hepatic One patient with cutaneous melanoma and multiple liver metas- infusion system had previously been implanted was treated sep- tases (diameter ≤ 20 mm) had a tumor response (tumor volume arately through the left and right hepatic arteries because of the decrease 95 %) after one chemosaturation-PHP treatment presence of new tiny vessels from a previously coiled GDA, feed- (●" Fig. 2). The residual tumor was treated successfully with ve- ing the duodenum and pancreas. murafenib plus laser-induced thermotherapy (LITT). On the MRI scan 3 months after chemosaturation-PHP, no more perfused tu- Hospitalization mor was visible. The patient remained tumor-free for 10 months; Frankfurt experience disease recurrence (> 50 hepatic lesions) was documented 13 After chemosaturation-PHP, patients were kept in the intensive months after treatment. Another patient with multiple liver me- care unit (ICU) or recovery room and transferred to the ward tastases from ocular melanoma had a partial response (●" Fig. 3). within a maximum of 24 hours when no complications occurred. Stable disease (tumor volume decrease < 5 %) was observed in a Patients remained in the hospital for 4 – 8 days. patient with hepatic and bone metastases from breast cancer. Two months after chemosaturation-PHP, the follow-up MRI scan This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.

Fig. 2 Magnetic resonance images (T2 haste) of a patient with diffuse liver Abb. 2 MRT-Bilder (T2 haste) eines Patienten mit diffusen Lebermetasta- metastases from malignant melanoma a, b and the corresponding images sen eines malignen Melanoms a, b und die entsprechenden Bilder 4 Wo- taken 4 weeks after chemosaturation-PHP c, d with significant reduction in chen nach der Chemosaturation-PHP c, d mit deutlicher Größenreduktion the size of the metastases (partial response). After two treatments with la- der Metastasen (partielle Remission). Nach zweimaliger Behandlung des ser-induced thermotherapy, full hepatic remission, which lasted for 10 Restbefundes mittels laserinduzierter Thermotherapie zeigte sich eine he- months, occurred. Regional disease recurrence (> 50 single lesions) patische Vollremission, welche für 10 Monate andauerte. Dann entwickelte was documented 13 months after treatment e, f. der Patient 13 Monate nach der Behandlung ein regionales Rezidiv mit mehr als 50 Einzelläsionen e, f.

Vogl TJ et al. Chemosaturation with Percutaneous … Fortschr Röntgenstr 2014; 186: 937–944 Interventional Radiology 941

Fig. 3 Magnetic resonance images (T2 haste) pre- treatment a and 8 weeks post-treatment b of a pa- tient with metastatic ocular melanoma. The liver metastases showed a reduction in size of over 30 % (partial response).

Abb. 3 MRT-Bilder (T2 haste) vor a und 8 Wochen nach der Behandlung b eines Patienten mit metas- tasiertem Aderhautmelanom. Die Lebermetastasen zeigen eine Größenreduktion von über 30 % (par- tielle Remission). This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.

Fig. 4 Magnetic resonance images (T2 haste) of the liver of a patient with Abb. 4 MRT-Bilder (T2 haste) einer Leber von einem Patienten, der mittels ocular melanoma who was treated with chemosaturation-PHP with mel- Chemosaturation-PHP mit Melphalan behandelt wurde a–c. 4 Wochen phalan a–c. Stable disease of diffuse small metastases was observed nach der Behandlung wurde eine stabile Situation diffuser kleiner Metasta- 8 weeks after treatment d–f. Without any further therapy, the patient sen beobachtet d–f. Ohne weitere Therapiemaßnahmen wies der Patient had a partial response in the liver 6 months after chemosaturation-PHP g–i. 6 Monate nach der Chemosaturation-PHP eine partielle Remission in der Leber auf g–i.

showed disease progression in the liver. Ten months after treat- sponse which is still ongoing 2 months later (●" Fig. 4). In another ment, the patient died from progression of liver metastases. The patient with bilobar disease, melphalan was infused only into the remaining 3 patients with stable disease had ocular melanoma. right hepatic artery, and complete hepatic perfusion was not Two months after chemosaturation-PHP, stable disease was achieved. Tumor response was evaluated only in the perfused maintained in 1 patient. Without any further therapy, the fol- right hepatic lobe (●" Fig. 5). The first follow-up after treatment low-up 6 months post-treatment imaging showed a partial re- showed stable disease in the right liver lobe and disease progres-

Vogl TJ et al. Chemosaturation with Percutaneous … Fortschr Röntgenstr 2014; 186: 937–944 942 Interventional Radiology

Fig. 5 Magnetic resonance images (MRI) of a 64-year-old man with metastatic ocular melanoma a, b. Due to vascular abnormalities, only the right liver lobe was treated with a single chemosatura- tion-PHP treatment. The postoperative MRI scans c, d showed stable disease in the right lobe, whereas disease progression occurred in the left lobe. (a, c: T2-haste, b, d: T2-TSE).

Abb. 5 MRT-Aufnahmen eines 64-jährigen Mannes mit metastasiertem Aderhautmelanom a, b. Aufgrund vaskulärer Anomalien wurde nur der rechte Leberlappen in einer einzelnen CS-PHP- Behandlung therapiert. Die postoperativen MRT- Aufnahmen c, d zeigten „stable disease“–Status im rechten Leberlappen bei gleichzeitigem Krank- heitsfortschritt im linken Leberlappen. (a, c: T2- haste, b, d: T2-TSE).

sion in the untreated left lobe. The left lobe was subsequently had a mixed response after the first perfusion, due to the growth treated with TACE. However, disease progression in both liver of some lesions deep within segment 7, located under the dome. lobes was documented after 2 months and the patient died A phrenic artery feeding that area was occluded immediately be- shortly afterwards. In the other patient with metastatic ocular fore the second session and a partial response was documented melanoma, the left and right hepatic arteries were cannulated in all target liver lesions (tumor volume reduction > 40 %). Unfor- separately and the right liver lobe was perfused with 2/3 of the tunately, this patient subsequently developed multiple lung me- melphalan dose and the left one with the remaining 1/3. One tastases. month after chemosaturation-PHP, stable disease was seen in the right liver lobe, whereas disease progression occurred in the Toxicity left lobe. A follow-up scan 2 months post-treatment showed pro- Frankfurt experience gressive disease in the whole liver. The patient died 6 months In the 2 patients treated with the first-generation filter, both de- after treatment from massive progression of liver metastases. veloped grade 1/2 increases in liver transaminases in the first 2 days after chemosaturation-PHP; all values returned to baseline Milan experience levels within 1 week. Grade 1/2 fatigue, nausea and fever within Of the 6 patients evaluable for response, 1 patient with biliary the first 5 days after treatment were also reported. In both pa- tract achieved complete remission 6 months tients, grade 1/2 bone marrow suppression was observed initi- after chemosaturation-PHP. 2 patients with ocular melanoma ally. However, nadir blood cell counts were reached after 11 – 12 had partial responses (MRI and PET) after the first of 3 sessions days and both patients were hospitalized for 8 – 9 days because of chemosaturation-PHP. One of these patients remained stable of grade 3/4 pancytopenia. Grade 3/4 leukocytopenia was treat- with minimal disease after 2 further sessions (MRI/PET-negative ed with filgrastim for 8 – 9 days in both patients, and antibiotics but some small nodules visible). The other patient showed dis- were required in 1 patient. Platelet transfusions were required

ease progression in the liver and bones after the third procedure, for grade 3/4 thrombocytopenia (n = 2), and 2 units of packed This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited. 1 year after the first session. One patient with cutaneous melano- red blood cells (RBC) were required for grade 3 anemia (n = 1). ma showed a partial response (MRI and PET) after the first che- Of the 4 patients treated with the new filter system, 1 patient de- mosaturation-PHP session. The patient then received 4 courses veloped grade 1/2 pancytopenia after treatment which worsened of ipilimumab and recently had a second chemosaturation-PHP to grade 3/4 pancytopenia after 1 week necessitating hospitaliza- session. One patient with gastric cancer had a mixed response tion for 6 days. However, only the leukocytopenia required treat- due to parasitic hepatic arteries that were not included during ment (filgrastim for 3 days plus an antibiotic). The second patient chemoperfusion. He had 3 consecutive liver resections before received a lower dose of melphalan and, with the exception of chemoperfusion. A huge portion of liver parenchyma, close to grade 1 anemia, did not experience any systemic adverse events. the previous resections, was fed by several arteries from the in- The last 2 patients both developed grade 1 anemia, grade 3/4 leu- tercostal and mammary arteries. Lesions located within the kocytopenia and grade 2 or 4 thrombocytopenia and were treat- non-perfused hepatic parenchyma grew after treatment, while ed as outpatients with platelet concentrates. Granulocyte colony- metastases within the chemoperfused parenchyma showed a stimulating factor (G-CSF) was given for 4 days. No significant in- partial response. Because it was not possible to manage this kind creases in transaminases occurred. of hepatic vascular supply, a second chemosaturation-PHP ses- A premenopausal patient developed vaginal bleeding after sys- sion was not scheduled. One patient with cutaneous melanoma temic heparinization. Treatment was stopped before melphalan

Vogl TJ et al. Chemosaturation with Percutaneous … Fortschr Röntgenstr 2014; 186: 937–944 Interventional Radiology 943

administration. A gynecological examination showed that the had no other therapeutic options (BRAF and C-kit wild type). event was most likely caused by heparin-induced bleeding of The last melanoma patient died after 9 months. These observa- the endometrium. The patient recovered without sequelae. tions compare favorably with the shorter median overall survi- val times typically reported in patients with liver metastases Milan experience from cutaneous [26] or ocular melanoma [27, 28]. In addition, Four patients were treated with the first-generation filter and all 5 other patients with ocular melanoma, gastric cancer or breast of them required multiple platelet and/or RBC transfusions after cancer experienced clear clinical benefit from the procedure, the procedure. One patient had grade 4 leukocytopenia, grade 3 opening up the possibility of combining chemosaturation-PHP anemia and grade 4 thrombocytopenia; the patient developed with other systemic treatments. febrile neutropenia was admitted to the hospital to receive intra- The toxicities observed in our patients were consistent with the venous antibiotics and other support therapies. Three patients profile of events documented in clinical trials [9, 11]. Mild to developed grade 4 leukocytopenia, grade 2/3 anemia and grade moderate filter-related toxicity, i. e. thrombocytopenia and ane- 1 or 4 thrombocytopenia which required transfusions without mia resulting from the removal of platelets and RBC by the hemo- hospitalization. All of these patients had grade 2/3 fatigue and re- filtration system, was observed immediately after the procedure. ceived filgrastim for at least 10 days. 3 of these 4 patients under- Only patients treated with the first-generation filter needed pla- went 1 – 2 more sessions of chemoperfusion with the second- telet and RBC transfusions. Persistent and more severe melpha- generation filter. These procedures were associated with mark- lan-related pancytopenia tended to emerge later. Although these edly reduced hematological toxicity (grade 1 anemia) without events were generally grade 3/4 in severity, they were predict- any need for blood or platelet transfusions; G-CSF was required able and were managed effectively in all patients with supportive for 4 days only. All 3 patients experienced milder fatigue (grade measures. Transient increases in liver transaminases (grade 1/2) 1). Of the 3 patients treated with the second-generation filter were also observed during the post-procedural period in some only, a similar toxicity profile was evident in 1 patient, while an- patients resulting from procedural manipulation of the liver rath- other patient developed febrile pancytopenia requiring transfu- er than melphalan. In all cases, values returned to baseline levels sions because of a hepatic vascular shunt that avoided drug filter- within 1 week. Other systemic events were mild (nausea, vomit- ing. No patient had significant liver dysfunction, and grade 1 – 2 ing, fever), apart from fatigue which significantly influenced pa- liver toxicity related to drug exposure was not influenced by the tient performance after treatment with the first-generation filter. type of filter used. The remaining patient, who had a large fast- Two events were attributed to heparin which is routinely admi- growing liver from ocular melanoma, died of a retro- nistered during chemosaturation-PHP to facilitate extracorporeal peritoneal giant hematoma 30 hours after chemosaturation-PHP. blood flow and hemofiltration. In one case the procedure was A post-mortem necropsy revealed multisite vascular bleeding stopped prematurely due to vaginal bleeding, and in another with no damage to the inner surface of the abdominal veins and case a perioperative death resulting from a retroperitoneal he- arteries. This unusual complication was most likely related to he- matoma occurred. parin which was needed for extracorporeal circulation. In April 2012, a second-generation high-efficiency filter (98 % bench-testing efficiency; Delcath Systems Inc., data on file) for use with the chemosaturation-PHP delivery system became com- Discussion mercially available in Europe. In Milan, three patients were treat- ! ed initially with the first-generation filter and then switched to Chemosaturation-PHP is a minimally invasive, repeatable tech- the new filter for repeated treatments, providing a unique oppor- nique which delivers high doses of chemotherapy directly to tu- tunity to compare toxicities with the two systems. In these pa- mors in the liver while limiting systemic toxicity through hemo- tients, toxicity with the new filter system was less severe and pa- filtration of the hepatic venous blood. Recently completed clinical tients required fewer supportive measures (i. e., no transfusions, trials performed in the US confirm the efficacy of chemosatura- shorter courses of colony-stimulating factors, reduced fatigue). tion-PHP delivery of melphalan in the treatment of patients Further studies are required to confirm this positive trend. A with hepatic metastases from melanoma [12, 13] and neuroen- phase III study to evaluate chemosaturation-PHP in the treat- docrine tumors [10, 11]. ment of is planned in Frankfurt, and Our experience with chemosaturation-PHP at the University the new filter system will also be investigated further.

Hospital Frankfurt and the European Institute of Oncology is In summary, chemosaturation-PHP for the treatment of non-re- This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited. supportive of the findings of the clinical trial program, although sectable liver metastases is a feasible treatment option when per- it is too early to make definitive statements about tumor re- formed by an experienced multi-disciplinary team. Hematologi- sponse rates and patient survival. Still, the data offer a number cal events, which are the predominant toxicities associated with of insights. A complete response was documented in 1 patient chemosaturation-PHP, are predictable and manageable with ap- with cholangiocarcinoma and partial responses were observed propriate supportive care. Based on our findings, we believe that in a further 6 patients with ocular or cutaneous melanoma, giv- chemosaturation-PHP is a promising technique for patients for ing an overall response rate of 50 % in the total patient sample whom there are no effective treatments. (n = 14). Of note, one of the melanoma patients became tumor- free after further treatment with LITT plus vemurafenib. The re- sponse lasted for 10 months before a recurrence in the liver was detected. At the time of writing, this patient is still alive with progression of liver metastases. 4 of the other 5 melanoma pa- tients are also alive 7 to 18 months after their first chemosa- turation-PHP session; all of them had disease restricted to the liver but were in progression after previous treatments and

Vogl TJ et al. Chemosaturation with Percutaneous … Fortschr Röntgenstr 2014; 186: 937–944 944 Interventional Radiology

Affiliations tastatic ocular or cutaneous melanoma. J Clin Oncol 2010; 28 (suppl; abstr LBA8512) 1 Department of Diagnostic and Interventional Radiology, University Hospital of Frankfurt 13 Pingpank JF, Hughes M, Alexander HR et al. Percutaneous hepatic per- 2 Department of Dermatology, Venerology and Allergology, University Hospi- fusion (PHP) vs. best alternative care (BAC) for patients (pts) with tal of Frankfurt melanoma liver metastases – efficacy update of the Phase 3 Trial 3 Department of Medicine I, University Hospital of Frankfurt (NCT00324727). Eur J Cancer 2011; 47: S653 (Suppl 1, abstr 9304) 4 Unit of Interventional Radiology, European Institute of Oncology, Milan 14 Lazarus HM, Herzig RH, Graham-Pole J et al. Intensive melphalan che- 5 Department of Anesthesiology, Intensive Care Medicine and Pain Therapy, motherapy and cryopreserved autologous bone marrow transplanta- University Hospital of Frankfurt tion for the treatment of refractory cancer. J Clin Oncol 1983; 1: 359– 6 Translational Research on Melanoma, European Institute of Oncology, Milan 367 15 Leff RS, Thompson JM, Johnson DB et al. Phase II trial of high-dose mel- phalan and autologous bone marrow transplantation for metastatic colon carcinoma. J Clin Oncol 1986; 4: 1586–1591 16 Alexander HR, Libutti SK, Bartlett DL et al. A phase I-II study of isolated References hepatic perfusion using melphalan with or without tumor necrosis 01 Breedis C, Young G. The blood supply of neoplasms in the liver. Am J Pa- factor for patients with ocular melanoma metastatic to liver. Clin Can- thol 1954; 30: 969–977 cer Res 2000; 6: 3062–3070 02 Vogl TJ, Mack MG, Eichler K et al. Chemoperfusion und -embolisation 17 Alexander HR Jr, Libutti SK, Pingpank JF et al. Hyperthermic isolated he- von Lebermetastasen [Chemoperfusion and embolization in the treat- patic perfusion using melphalan for patients with ocular melanoma ment of liver metastases]. Fortschr Röntgenstr 2011; 183: 12–23 metastatic to liver. Clin Cancer Res 2003; 9: 6343 –6349 03 Beheshti MV, Denny DF Jr, Glickman MG et al. Percutaneous isolated liv- 18 Noter SL, Rothbarth J, Pijl ME et al. Isolated hepatic perfusion with high- er perfusion for treatment of hepatic malignancy: preliminary report. J dose melphalan for the treatment of uveal melanoma metastases con- Vasc Interv Radiol 1992; 3: 453–458 fined to the liver. Melanoma Res 2004; 14: 67–72 04 Ravikumar TS, Pizzorno G, Bodden W et al. Percutaneous hepatic vein 19 Rizell M, Mattson J, Cahlin C et al. Isolated hepatic perfusion for liver isolation and high-dose hepatic arterial infusion chemotherapy for un- metastases of malignant melanoma. Melanoma Res 2008; 18: 120– resectable liver tumors. J Clin Oncol 1994; 12: 2723 –2736 126 05 Curley SA, Newman RA, Dougherty TB et al. Complete hepatic venous 20 van Iersel LB, Gelderblom H, Vahrmeijer AL et al. Isolated hepatic mel- isolation and extracorporeal chemofiltration as treatment for human phalan perfusion of colorectal liver metastases: outcome and prognos- hepatocellular carcinoma: a phase I study. Ann Surg Oncol 1994; 1: tic factors in 154 patients. Ann Oncol 2008; 19: 1127 –1134 389–399 21 Alexander HR Jr, Bartlett DL, Libutti SK et al. Analysis of factors associat- 06 Hwu WJ, Salem RR, Pollak J et al. A clinical-pharmacological evaluation ed with outcome in patients undergoing isolated hepatic perfusion for of percutaneous isolated hepatic infusion of doxorubicin in patients unresectable liver metastases from colorectal center. Ann Surg Oncol with unresectable liver tumors. Oncol Res 1999; 11: 529–537 2009; 16: 1852–1859 07 van Etten B, Brunstein F, van IJken MG et al. Isolated hypoxic hepatic 22 Rothbarth J, Pijl ME, Vahrmeijer AL et al. Isolated hepatic perfusion with perfusion with orthograde or retrograde flow in patients with irre- high-dose melphalan for the treatment of colorectal metastasis con- sectable liver metastases using percutaneous balloon catheter tech- fined to the liver. Br J Surg 2003; 90: 1391 –1397 niques: a phase I and II study. Ann Surg Oncol 2004; 11: 598–605 23 Feldman ED, Wu PC, Beresneva T et al. Treatment of patients with unre- 08 Verhoef C, de Wilt JH, Brunstein F et al. Isolated hypoxic hepatic perfu- sectable primary hepatic malignancies using hyperthermic isolated sion with retrograde outflow in patients with irresectable liver metas- hepatic perfusion. J Gastrointest Surg 2004; 8: 200–207 tases; a new simplified technique in isolated hepatic perfusion. Ann 24 Grover AC, Libutti SK, Pingpank JF et al. Isolated hepatic perfusion for Surg Oncol 2008; 15: 1367–1374 the treatment of patients with advanced liver metastases from pancre- 09 Pingpank JF, Libutti SK, Chang R et al. Phase I study of hepatic arterial atic and gastrointestinal neuroendocrine neoplasms. Surgery 2004; melphalan infusion and hepatic venous hemofiltration using percuta- 136: 1176–1182 neously placed catheters in patients with unresectable hepatic malig- 25 Therasse P, Arbuck SG, Eisenhauer EA et al. New guidelines to evaluate nancies. J Clin Oncol 2005; 23: 3465 –3474 the response to treatment in solid tumors. European Organization for 10 Pingpank JF, Royal RE, Kammula US et al. High dose intra-arterial Research and Treatment of Cancer, National Cancer Institute of the melphalan delivered via percutaneous hepatic perfusion (PHP) for pa- United States, National Cancer Institute of Canada. J Natl Cancer Inst tients with unresectable hepatic metastases from primary neuroendo- 2000; 92: 205–216 crine tumors. Fort Lauderdale Florida, USA: American Hepato-Pan- 26 Rose DM, Essner R, Hughes TM et al. Surgical resection for metastatic creato-Biliary Association Annual Meeting; 2008 melanoma to the liver: the John Wayne Cancer Institute and Sydney 11 Pingpank J, Royal RE, Kammula US et al. Percutaneous hepatic perfu- Melanoma Unit experience. Arch Surg 2001; 136: 950–955 sion (PHP) with melphalan for patients with unresectable liver metas- 27 Papastefanou VP, Cohen VML. Uveal melanoma. J Skin Cancer 2011: tases of neuroendocrine tumors (MNET) – NCT000960083. Eur J Can- 573974, Epub 2011 Jun 30 cer 2011; 47: S478 (Suppl. 1, abstr 6621) 28 Feldman ED, Pingpank JF, Alexander HR Jr. Regional treatment options 12 Pingpank JF, Hughes MS, Faries MB et al. A phase III random assignment for patients with ocular melanoma metastatic to the liver. Ann Surg trial comparing percutaneous hepatic perfusion with melphalan (PHP- Oncol 2004; 11: 290–297

mel) to standard of care for patients with hepatic metastases from me- This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.

Vogl TJ et al. Chemosaturation with Percutaneous … Fortschr Röntgenstr 2014; 186: 937–944